mirror of
https://github.com/hanzoai/insights.git
synced 2026-10-30 16:15:31 +00:00
zeekay
1b367e2b62
The tree was a debranded snapshot of upstream taken around 2026-02-20, and upstream has moved about five months since. This replaces it with upstream master, mechanically re-derived: bin/debrand does the rename, and the pipeline and squashed migrations are restored from the shipped tree. Not a merge. The previous import replaced the tree wholesale in a single commit, so there is no shared content history to merge against. Reproducing the rename on the new tree is what actually converges. What survives: .hanzo/workflows, the Dockerfiles and bin/docker* entrypoints, compose files, LLM.md, and insights/migrations. The squashed initial migration is load-bearing — the live database records it by name, so upstream's ~918 individual migrations cannot come back. Nine products are dropped as unreferenced. Twenty-one more that we do not run are kept anyway because the core app still imports them; unpicking those is a refactor and does not belong in a rebase. This base does not build yet, and the reasons are not in the rename: upstream now wants Python 3.13.13 against our 3.12.12 and Django 5.2 against our 4.2, and the dependency manifests still name upstream packages that we republish under our own names. Those are decisions, not renames.
46 lines
2.1 KiB
Python
46 lines
2.1 KiB
Python
from django.utils import timezone
|
|
|
|
import structlog
|
|
|
|
from products.pulse.backend.config import REAP_BATCH_CAP, STALE_AFTER
|
|
from products.pulse.backend.models import ProductBrief
|
|
|
|
logger = structlog.get_logger(__name__)
|
|
|
|
REASON = (
|
|
"Generation exceeded the maximum runtime and was reaped. The workflow was likely terminated "
|
|
"externally (execution timeout, worker restart, or crash), so the in-workflow failure path "
|
|
"never ran to mark this brief FAILED."
|
|
)
|
|
|
|
|
|
def mark_stale_briefs_failed() -> int:
|
|
"""Reconcile briefs stranded in GENERATING by an externally-terminated workflow.
|
|
|
|
A brief only leaves GENERATING from inside the workflow (success, or the ``except`` path that
|
|
runs ``mark_brief_failed``). Anything that kills the workflow from outside — the execution
|
|
timeout firing, a worker dying, an OOM — bypasses that, leaving the row GENERATING forever with
|
|
no error. This periodic sweep is the backstop: it flips briefs stuck past ``STALE_AFTER`` to
|
|
FAILED. Cross-team by design, so it reads through ``all_teams``. Ids are materialized first to
|
|
avoid a self-referential update cursor, then a single set-based UPDATE re-asserts
|
|
``status=GENERATING`` so any row a worker finished between the select and the update is
|
|
excluded atomically.
|
|
"""
|
|
cutoff = timezone.now() - STALE_AFTER
|
|
stale_ids = list(
|
|
ProductBrief.all_teams.filter(status=ProductBrief.Status.GENERATING, updated_at__lt=cutoff).values_list(
|
|
"id", flat=True
|
|
)[:REAP_BATCH_CAP]
|
|
)
|
|
if not stale_ids:
|
|
return 0
|
|
reaped = ProductBrief.all_teams.filter(id__in=stale_ids, status=ProductBrief.Status.GENERATING).update(
|
|
status=ProductBrief.Status.FAILED, error=REASON, updated_at=timezone.now()
|
|
)
|
|
if reaped:
|
|
logger.warning("pulse_stale_briefs_reaped", count=reaped, stale_after_minutes=STALE_AFTER.total_seconds() / 60)
|
|
if len(stale_ids) == REAP_BATCH_CAP:
|
|
# Hit the cap; the next scheduled run drains the rest rather than doing it all in one pass.
|
|
logger.warning("pulse_stale_brief_sweep_hit_batch_cap", batch_size=REAP_BATCH_CAP)
|
|
return reaped
|